Model-assisted process design for better evaluation and scaling up of continuous downstream bioprocessing.

Continuous process is a promising alternative for tradition batch process in biomanufacturing, which has higher productivity and lower material consumption. However, despite the maturation of necessary technologies for continuous process, there are few discussion about optimization of full continuous process. One possible reason is that the continuous process is such a complex and interacted process that the traditional experiment-based optimization approach is not sufficient anymore. To address that problem, the process simulation tool SuperPro Designer and continuous capture chromatography model were integrated into a model-assisted design platform for better development of continuous process. The influences of different continuous capture operation modes and sub-lot number under various upstream conditions were analyzed for pilot-scale production. The best combination of operation mode and sub-lot number were determined for the highest equipment utilization without any conflict. After the process optimization, the equipment utilization of continuous process was significantly improved to 27.3%. Then, a pilot-scale case study was carried out to validate the proposed approach. The results showed that the scaling up and process design of continuous process were successful. No time conflict and process failure happened and the final product met the release specification. Finally, the cost of goods was evaluated with SuperPro Designer, and the results showed a 17.4% cost reduction for optimized continuous downstream process compared to batch process, which is promoting for the future industrial applications.

Converting a mAb downstream process from batch to continuous using process modeling and process analytical technology.

The biopharmaceutical market is driving the revolution from traditional batch processes to continuous manufacturing for higher productivity and lower costs. In this work, a batch mAb downstream process has been converted into an integrated continuous process with the combination of multiple techniques. For process intensification, two batch mode unit operations (protein A capture chromatography, ultrafiltration/diafiltration) were converted into continuous ones; for continuity, surge tanks were used between adjacent steps, and level signals were used to trigger process start or stop, forming a holistic continuous process. For process automation, manual operations (e.g., pH and conductivity adjustment) were changed into automatic operation and load mass was controlled with process analytical technology (PAT). A model-based simulation was applied to estimate the loading conditions for the continuous capture process, resulting in 21% resin capacity utilization and 28% productivity improvement as compared to the batch process. Automatic load mass control of cation exchange chromatography (CEX) was achieved through a customized in-line protein quantity monitoring system, with a difference of less than 1.3% as compared to off-line analysis. Total process time was shortened from 4 days (batch process) to less than 24 hours using the continuous downstream process with the overall productivity of 23.8 g mAb per day for the bench-scale system. Comparable yield and quality data were obtained in three test runs, indicating a successful conversion from a batch process to a continuous process. The insight of this work could be a reference to other similar situations.

Process development and optimization of continuous capture with three-column periodic counter-current chromatography.

Continuous capture with affinity chromatography is one of the most important units for continuous manufacturing of monoclonal antibody (mAb). Due to the complexity of three-column periodic counter-current chromatography (3C-PCC), three approaches (experimental, model-based, and simplified approaches) were studied for process development and optimization. The effects of residence time for interconnected load (RT C ), breakthrough percentage of the first column for interconnected load (s) and feed protein concentration (c 0 ) on productivity and capacity utilization were focused. The model-based approach was found superior to the experimental approach in process optimization and evaluation. Two phases of productivity were observed and the optimal RT C for the maximum productivity was located at the boundary of the two phases. The comprehensive effects of the operating parameters (RT C , s, and c 0 ) were evaluated by the model-based approach, and the operation space was predicted. The best performance of 34.5 g/L/h productivity and 97.6% capacity utilization were attained for MabSelect SuRe LX resin under 5 g/L concentration at RT C = 2.8 min and s = 87.5%. Moreover, a simplified approach was suggested to obtain the optimal RT C for the maximum productivity. The results demonstrated that model-assisted tools are useful to determine the optimum conditions for 3C-PCC continuous capture with high productivity and capacity utilization.

Improving catalytic performance of an arylacetonitrilase by semirational engineering.

Arylacetonitrilases have been widely acknowledged as important alternatives to chemical catalysts for synthesizing optically pure 2-hydroxyphenylacetic acids from nitriles. In this work, two residues (Thr132 and Ser190) located at the catalytic tunnel in the active site of an arylacetonitrilase nitA from uncultured organisms were mutated separately by site-directed mutagenesis. Ser190 was demonstrated to be the critical position which has a greater influence on arylacetonitrilase nitA activity than Thr132. The replacement of serine at position 190 with glycine increases its activity toward mandelonitrile and (o, m, p)-chloromandelonitrile, whereas replacing it with leucine abolished its activity. The best mutant S190G exhibited threefold higher specific activity toward mandelonitrile compared with that of wild-type nitA, which rendered it promising for industrial application. Homology modeling and molecular docking experiments were in agreement with the kinetic assays and support the improved catalytic performance.